@@i The American Antitrust Institute
Albert A. Foer President The American Antitrust Institute “Can Antitrust Save Academic Publishing?”
American Library Association Annual Meeting Orlando, Florida
June 28, 2004 (revised July 20, 2004)
2919 ELLICOTT ST, NW • WASHINGTON, DC 20008 PHONE: 202-276-6002 • FAX: 202-966-8711 •
[email protected] www.antitrustinstitute.org
In thinking about the present state of academic publishing, four observations seem to be of particular pertinence. First, the industry has been marked by a large number of mergers, leading it to become more and more concentrated. Second, prices for academic publications have been advancing far faster than in the economy generally. Third, the Internet has revolutionized the ways in which academic information is being disseminated. And fourth, a new practice of so-called Big Deal Bundling has emerged as a way of sheltering some publishers from the “storm of creative destruction” threatened by the Internet. Before pondering the implications of these observations and exploring what role the antitrust laws might play, let’s take a moment to document the four observations . Observation 1. The industry has been marked by a large number of mergers, leading it to become more and more concentrated. The industry we are talking about publishes academic journals, which include both for-profit and non-profit publications. We have to be careful in talking about the “industry” because sometimes we may be referring to the entire industry and sometimes we may be referring to smaller markets made up of segments of the industry, such as the legal journals market or the geological journals market. The literature on this industry often refers, for example, to the STM (scientific, technical, and medical) part of the market. When we say that the industry is becoming more concentrated, this does not necessarily imply that every smaller market contained within the industry is also becoming more concentrated, although the two seem to be related. Right now, we will focus on the overall picture. I do not have a complete economic portrait of the industry to present. Apparently, there are something like 30,000 academic journals in the world and over 8,000 peerreviewed journals in various fields.1 The peer-reviewed journals, being the most prestigious among academics, are the ones of most importance for libraries. While 8,000 1 Mark J. McCabe, “Journal Pricing and Mergers: A Portfolio Approach,” 92 The American Economic Review 259, 262 (March 2002).
2
may seem a large number, it is clear that a substantially increasing percentage of journals have come under the ownership of only a small number of companies. That is to say, the industry is becoming more concentrated. For a portrait of twelve leading academic publishers and their individual histories of mergers and acquisitions, one can turn to Mary H. Munroe’s survey,2 which is guaranteed to set your head spinning. Significant merger activity in academic publishing has occurred since 1990. It is not just that small publishers are being taken over by larger ones; the large ones are also gobbling up each other. McCabe, looking at the STM market, noted in 2002 that at least five major commercial publishers had been acquired by competitors since 1997. Since the late 1980’s, Wolters Kluwer alone has acquired more than 300 companies. When ReedElsevier purchased Harcourt General in 2001, this brought Reed-Elsevier’s total of ISIranked biomedical titles alone to 409.3 According to a forthcoming paper by Edlin and Rubenfeld, two distinguished economists, in 2001, measured by revenue, Elsevier Science had a 16.0% industry share; Kluwer, 8.2%; and Thomson-Scientific & Healthcare, 7.5%. And the top ten STM publishers (which includes the National Library of Medicine) accounted for 63.4% of the industry. These shares include both non-profit and for-profit journals. Restricting the counting to the commercial publishers segment of the STM industry, Elsevier Science’s share is 22.9%, Kluwer’s is 11.7%, and Thomson’s is 10.7%.4 Assuming these figures are accurate, they do not indicate a highly concentrated industry by current antitrust standards, even if we focus on the commercial sector alone. Moreover, this data uses revenues as the measure of concentration. If one were to look at the number of journals rather than their revenue production, one method for examining 2
Mary H. Munroe, “The Academic Publishing Industry: A Story of Merger and Acquisition,” http://www.niulib.niu.edu/publishers/ (visited 6/4/2004). 3
ISI, the Institute for Scientific Information, indexes and tracks citations for approximately 8,500 of the most prestigious, high impact research journals in the world. 4
Aaron Edlin and Daniel Rubinfeld, “Exclusion or Efficient Pricing? The ‘Big Deal’ Bundling of Academic Journals, forthcoming, Antitrust L.J. (2004).
3
diversity, the odds are that there would be a lower measure of concentration. On the other hand, if one looked at prestige (which helps generate higher revenues), it may be more highly concentrated. The overall situation may be characterized generally not as one of monopoly or tight oligopoly, but of relatively loose oligopoly becoming a tighter oligopoly. Finally, as I will discuss later on, there are several possible ways to define the relevant antitrust market, and one cannot place too much weight on market shares until one has a consensus on the market definition. It would be useful to have current concentration data, measured in various ways, and comparable to the past, in order to identify the trends more precisely, but the overall picture seems undeniably to be one of rapidly increasing concentration. It is likely, though I do not have the data to say this with certainty, that concentration in some more narrowly defined markets (e.g. chemistry, or medicine, or law; or, even more narrowly, biochemistry, neurology, or antitrust law), is considerably higher.
Observation 2. Prices have been advancing far faster than in the economy generally. Subscription prices for academic journals have been moving upward at a much faster rate than inflation, even faster than such inflationary benchmarks as health care or college tuition. “The price of library subscriptions to periodicals in law, medicine, and physics and chemistry rose by 205 percent, 479 percent and 615 percent between 1984 and 2001, a period when overall price increases as reflected by CPI was 70 percent.”5 Between 1991 and 2000, library subscriptions to STM journals increased in price 158 percent, over six times the inflation rate, while legal serial publications increased 103 percent, over four times the inflation rate.6 It is also noteworthy that the price increases of
5
Id.
6 Thomas M. Susman, David J. Carter, “Publisher Mergers: A Consumer-Based Approach to Antitrust Analysis,” at 3, prepared for the Information Access Alliance, June 2003, and available on the Alliance’s website at http://www.informationaccess.org.
4
commercial publishers have been significantly higher than those of nonprofit publishers.7 The cause cannot be laid off on the increased cost of paper, which went up by only 12 percent from 1991 to 2000.8
Observation 3. The Internet has revolutionized the ways in which academic information can be disseminated. In the last several years, electronic publishing has become widespread. It is now feasible to make the content of academic journals available on the Internet, with dramatically lower distribution costs. It is also possible that electronic publishing also reduces the so-called first-copy costs (recruiting writers, judging, reviewing, editing, copy editing, and typesetting articles). Whether electronic publishing will lead to a substantial number of new entrants into academic publishing remains to be seen, because there are many other structural barriers to entry, including—most importantly—the journal’s reputation. Other barriers are network effects, coordination, the existing stock of journals held by incumbents, and the high switching costs for libraries. What is already clear, however, is that the threat of electronic published has forced the industry to reconsider its pricing strategies.
Observation 4. A new practice of so-called Big Deal Bundling has emerged as a way of sheltering some publishers from the “storm of creative destruction” threatened by the Internet. All of the major commercial publishers have changed their pricing strategies, introducing what many librarians call the “Big Deal.” Presumably, this occurred without collusion, since a horizontal agreement on pricing by competitors would constitute the 7
Edlin and Rubinfeld, op. cit., cite data showing that the average price per page of elite commercial journals is about 9 times as high as for non-profit journals.
5
ultimate antitrust sin and more or less automatically subject the publishers to huge potential damages and even criminal prosecution. (Of course this presumption of innocence could be overcome by evidence. Cartel behavior, unfortunately, is not rare. But I am not aware of any allegations that it an issue here.) The detailed terms of agreements vary from publisher to publisher and from subscriber to subscriber, but the essence of the “Big Deal” is the bundling together of print and digital representations of journals, with the library agreeing to keep paying for its print subscriptions (with rising prices) for a period of years and, paying a surcharge, gaining access to an electronic journal database. The point of the arrangement for the publisher is to keep the library from canceling its print subscriptions once it has access to the electronic publishing and this is accomplished by the pricing structure, so that any alternative to taking the Big Deal will be more expensive for the library. From the library’s perspective, gaining access to the electronic database is valuable. Possible Implications of These Observations An antitrust attorney naturally finds the four preceding observations to be inviting. In any industry where there is a clear merger wave, the question of whether any particular merger violates Section 7 of the Clayton Act is bound to arise. There is often a linkage between increased levels of industry concentration and higher prices, and so the apparent combination of many mergers and rapidly increasing prices raises the question of causation. The advent of the Internet adds another facet: in many industries, antitrust attorneys are familiar with efforts by a so-called legacy segment trying to choke off new entrants made possible by a new technology.9 And finally, the bundling of one product with another, sometimes called tying or full-line forcing, is a sticky antitrust phenomenon that has only recently been in the news with respect to a product as well-known as Scotch Tape.
8 9
Susman and Carter, op. cit. at 3. See Albert A. Foer, “E-Commerce Meets Antitrust: A Primer,” 20 J. Pub. Policy & Marketing 51 (2001).
6
In the remainder of this paper, we will provide the legal framework for, first, analyzing mergers in the academic publishing market, and, second, analyzing bundling arrangements. The mergers of concern in this market are characterized as horizontal. That is, they occur between firms operating at the same level of the industry: publisher beds down with publisher. You may remember “bundling” from your social history of colonialism in America as being somewhat similar, but with a board between the participants. In antitrust, however, bundling, is a vertical issue, not a horizontal one, relating to two levels in the supply chain: publisher selling to library. One of the questions we will address is whether there is some relationship between the horizontal and vertical issues.
The Merger Framework Historically, American merger law is mostly bound up with Section 7 of the Clayton Act, which focuses on conduct, the effect of which “may be substantially to lessen competition, or to tend to create a monopoly in any line of commerce.” 10 You’ll notice that this is forward-looking language. It reaches anticompetitive conduct in its incipiency, meaning that there must be a reasonable probability that the conduct challenged would, if undisturbed, mature into a restraint of trade. The Clayton Act covers both acquisitions and mergers, whether by stock or asset purchase, and I will use the word merger to cover any transaction that can be examined under the Clayton Act. The Clayton Act was passed in 1914 and although there have been a few statutory changes, the law on paper has not really changed in 90 years. One extremely important procedural change came in 1976 when the Hart-Scott-Rodino Premerger Notification Law was passed.11 This requires all mergers over a certain threshold size to be put on hold until either the Federal Trade Commission or the Department of Justice has a chance to take a look at information that must be provided by the merging parties. About 97% of 10
15 U.S.C. sec. 18.
11
15 U.S.C. sec 18A.
7
the time, the government allows pre-notified mergers to be consummated within 30 days, without further ado. The remaining mergers can be delayed while the enforcement agency makes what is called a “second request,” asking for and evaluating further information. If the agency believes there is something objectionable, it may seek a preliminary injunction in a federal district court. Typically, if it wins the injunction, the parties walk away from the merger, although they are entitled to a full hearing and trial. Usually the prospective cost and especially the delay of a trial are enough to deter the parties. The Hart-Scott-Rodino reform allows the agencies to stop mergers before they occur, rather than having to try to “unscramble the eggs” after the fact. This has revolutionized merger control and variations have been adopted all around the world. Part of the revolution is that the vast majority of mergers are now handled more as a matter of administration than of adjudication. The agencies have set forth detailed guidelines for how horizontal mergers should be analyzed and the courts, on those relatively rare occasions when they have the opportunity to speak about merger law, usually accept these guidelines.12 The guidelines become the basis for negotiations between the parties and the government, and these negotiations frequently result in a settlement order that eliminates the parts of the merger that were deemed anticompetitive. Between one and two percent of the mergers that get notified are either stopped or restructured as a result of antitrust review. Of course, an unknown number of other merger deals simply do not get made because of legal advice that the merger will not pass inspection. Many of the most troublesome mergers have been allowed to go forward, once certain competitive overlaps between the merging parties have been eliminated. As an example, when the Justice Department reviewed the purchase of West Publishing Company by Thomson Financial & Professional Publishing Group in 1996, it was concerned about instances of content overlap between individual titles that the two parties published. It conditioned approval of the merger on the divestiture of those assets where
12
U.S. Dept. of Justice and F.T.C., Horizontal Merger Guidelines (1992, as amended 1997), available at http://www.ftc.gov/bc/docs/horizmer.htm.
8
content overlapped. Although the government has the ability to stop a merger outright, the typical remedy is to fix the anticompetitive problem and then let the merger proceed if the parties are still interested. My description of the procedural framework is now almost complete. You are already aware that there are two federal agencies, the FTC and the DOJ, that divide up the merger workload but apply the same rules. The division of labor is usually based on prior experience, so as the West case suggests, future mergers involving academic publishing will most likely be handled by the DOJ, although mergers in which the Internet plays an important role may go to the FTC. You should also be aware that state attorneys general and private parties may also bring enforcement actions. States will generally limit themselves to situations where the anticompetitive impact is particularly important within the state’s borders. Private challenges to mergers have become relatively rare because of procedural impediments. And finally, as I hinted, there are now approximately one hundred countries with their own antitrust laws, and these can have an impact on large mergers involving international business such as may occur in academic publishing. Turning to substantive law, when a government evaluates a horizontal merger, it worries primarily about three types of anticompetitive effects. First, will the surviving firm have assembled the instruments of dominance? That is, will the merger create a monopoly, with all of the ills we associate with monopoly? Second, with fewer players in the market and uncertainty thereby reduced, will the market be more susceptible to collusion? And third, will the merged firm itself be able to raise prices unilaterally? The second category is referred to as coordinated effects and the third category as unilateral effects. The process of analysis is set forth in the Horizontal Merger Guidelines. First, they require the relevant market to be defined. Then the market is placed in one of three categories: unconcentrated, moderately concentrated, or highly concentrated. This is done by arithmetic manipulation of market share percentages. No merger in an unconcentrated
9
market will be challenged. In a moderately concentrated market, only large mergers leading to greater market concentration will trigger antitrust concern. And in a highly concentrated market, almost any merger will be scrutinized closely. After this structural triage is completed and a closer look is found to be warranted, the government will consider a variety of other market factors. For example, will postmerger market conditions be conducive to reaching terms of coordination and detecting and punishing deviation from these terms? Perhaps more relevant to the mergers in academic publishing, will the merged firm be able to raise prices unilaterally? In a market characterized by product differentiation, such as academic publishing, a merger that combines the first and second choices of many consumers may well lead to price rises to certain classes of customers, regardless of whether all competitors join the increase. Next, the government looks at conditions of entry. How easy is it for newcomers to enter the industry, such that any anticompetitive effects would be relieved in a timely manner? The general rule is that entry will be considered timely if it is likely to occur within two years. Before the Internet, the academic publishing industry had high barriers to entry primarily because of the difficulty in building a reputation for a new publication. Argument may be expected over the question of how much of a difference the Internet makes. Finally, the guidelines allow for recognition of efficiencies that may be created by the transaction. The efficiencies must be merger-specific, that is, unique to the particular transaction, not reasonably attainable by other means. The role of efficiencies remains controversial. They have not often been found to outweigh anticompetitive effects. 13
13
In terms of defenses to otherwise anticompetitive mergers, it should also be noted that a merger will not be challenged if the acquired firm offers a convincing “failing-firm defense.”
10
Market Definition and Academic Publishing In many antitrust cases, the most crucial determination is the definition of the relevant market. Market definition, a process I will argue in the face of scowling economists is more closely related to art than science, is accomplished by looking separately at supply and demand conditions and trying to ascertain how truly various products are viable substitutes for one another. When two of the three leading cruise companies merge, if the relevant market is deemed to be cruises, the merger would seem to be problematic. On the other hand, if the relevant market is vacations, including landbased vacations, then the cruise companies have only a small share of the market and their merger may seem to have an insignificant impact on competition. With academic publications, there is the similar challenge of figuring out how broad or how narrow the relevant market is. Here are three different possible approaches: (a) The market is very broadly inclusive. Academic publications compete for the reader’s dollar with periodicals and books of all sorts, possibly even with computers or chairs for a library, depending on whose perspective is taken. This broad definition is the equivalent of putting cruise vacations into the same market with land-based vacations. Most people would agree that it is too far-reaching. When a reader needs a book about economics, a Latin grammar just won’t do. (b) The market is a portfolio of academic publications that deal generally with similar subject matter, such as science, technology and medicine; or perhaps more narrowly, with legal publications. From a reader’s perspective, these categories, in the broader or narrower form, may still be too broad. If I need a law book about antitrust, a book about divorce law will not be a suitable substitute. Note, however, that we are looking at this from the perspective of the reader, the user of the publication. At least in the case of libraries, the library may have different criteria and may in fact make purchasing decisions on the basis of portfolios, where it is not individual titles that compete but
11
rather the buyer selects among more general categories that compete for the limited procurement dollars. Whether this important distinction should affect an antitrust court’s determination of a relevant market is an interesting question indeed. (c) The market is content-defined. This is the narrowest market definition. It recognizes a high degree of market differentiation. All law books are not the same. Indeed, all antitrust law books are not interchangeable. A book about merger law in the U.S. cannot be substituted for one about merger law in the U.K., not to mention a book about traditional tribal law written in Swahili. The same goes, of course, for journals, where neither journals nor the articles within them can be substituted precisely one for the other. Query, in an extreme case, can a title or journal be so indispensable as to constitute its own product market? Consistent with other mergers, the government, according to some who follow this industry closely, appears to have opted in academic publishing mergers for the third, content-driven market definition, although not to the extreme of saying that each title is a separate market. In a merger of two supermarket chains, the government typically allows the merger to go through after individual supermarket locations within a certain radius of one another are divested. In this way, it is said that competition has not been diminished by the merger because the only parts of the companies that were competing will still compete after the merger. In publishing, the practice seems to be that the merger is broken down into the smaller segments in which titles compete directly against one another, from the perspective of the reader. Once it is assured that competing titles in concentrated categories are divested, there is no further cause for concern. Or is there? The result of this standard approach is paradoxical. The same number of titles remain in the market, but the number of publishers is reduced. This may not make much difference if there are still many publishers left and they are very roughly of
12
similar size. But what happens in the market when there are only a small handful of dominant rivals? These dominant firms have much more power in the market than do the competitors in a fragmented market, and they have this power both on the selling side and the buying side. On the selling side, their power seems to turn on the way in which business is done with the various buyers of their product. Librarians, for instance, constitute a major class of buyers. They traditionally work within budgets that cover a portfolio of topics, with the goal of purchasing the journals that are most in demand by their constituents, i.e. the readers in general. If librarians make purchasing decisions based on portfolios rather than individual titles, it can be argued that the nature of competition in this industry is somewhat different than in other industries. That is, while a reader may not consider an antitrust journal to be a viable substitute for a divorce law journal, a law librarian, using different criteria within a limited budget, makes decisions that in effect consider all law titles to be competing against one another. Thus, a market definition that looks only to the reader may be inadequate to explain how competition actually works in the academic publishing industry. Similar considerations can be present in other markets where the actual buyer is different from the user in whose name the purchase is made, e.g., insured medical care. Good antitrust analysis is based on a realistic understanding of how a market operates. If this unique perspective on procurement is taken for the academic publishing business, then it is not sufficient to protect the number of titles; it might also be necessary to assure that there remain enough publishers with portfolios so that librarians will have a reasonable range of choice. One cannot jump at this conclusion without doing more homework. For example, is there a way that librarians can modify their behavior to adjust to a smaller number of sellers? If not, why not? As one part of this answer, we will turn to the question of bundling, but I want to make one other point first.
13
Buyer Power What about the enhanced buyer power that academic publishers attain when they merge? Publishers stand in the position of a buyer with respect to the academic content providers, i.e. the professors and others who write the articles. In effect, even if there is no cash payment that moves from publisher to writer, compensation takes the form of prestige and distribution. Writers need a reasonable range of publication choices when they decide what to write about, how to format their writing, and how to distribute their work. The publishers exercise huge influence not only on these factors but also on a writer’s decision of what content and format to pursue. A reduction in the number of choices for the writer, at least when the remaining number of competing publishers is becoming small, can be expected to shift the balance of power that shapes bargaining between writer and publisher. When writers are limited in their publication outlets, the market is skewed, but more importantly there are large negative implications for a society that believes in academic freedom and values free-ranging research and expression. These values tend not to be taken into account openly by today’s antitrust enforcers, who focus almost exclusively on whether a merger is likely to result in higher prices in the near term. There is, however, not full agreement on this within the antitrust community. For instance, many antitrust experts contend that consumer choice is also an objective of the antitrust laws. Such experts, probably representing a minority view at the present time, might argue that a merger of large academic publishers leading to further concentration of ownership and fewer decision makers regarding what will be published, is likely to create a situation in which consumer choice will be unduly limited by a merger, and they will argue that this is within the scope of the Clayton Act even if the price of a publication is not about to increase as a result of the merger.
14
The Bundling Framework In a typical “Big Deal” that is offered by an academic publisher to a library, the library enters into a long-term arrangement to get access to a large electronic library of journals at a substantial discount in exchange for a promise not to cut print subscriptions, even if the print subscriptions become more expensive over time.14 The Big Deal is structured so that it is by far less expensive than the other alternatives that are offered to the library. The competitive problem caused by the Big Deal grows out of the fact that once a library is committed to a Big Deal with a major publisher, it cannot save money by cancelling subscriptions to particular titles. Because libraries have limited budgets and the costs of titles keep escalating, this has the practical effect of foreclosing the library from purchasing titles from other publishers or, alternately stated, of foreclosing other sellers from selling to the libraries. In effect, the Big Deal creates a major new strategic entry barrier to entry into the journals market. Smaller publishers object because they are being denied access to the libraries. Librarians don’t like it because they are unable to save money and their ability to purchase the publications they want is being reduced. How shall antitrust analysis approach the Big Deal arrangement? Tying The Big Deal could be challenged under the antitrust laws as a form of what is called a tying arrangement15 or as monopoly maintenance. Under a tying arrangement, the seller of a product conditions the sale of one product upon the buyer’s agreement to
14
In my discussion of the “Big Deal,” I am relying heavily on Aaron Edlin and Daniel Rubinfeld, “Exclusion or Efficient Pricing? The ‘Big Deal’ Bundling of Academic Journals,” forthcoming, Antitrust Law Journal (2004).
15
A simpler form of tying may also occur within the academic publishing industry. Electronic publishers may price their offerings in such a way that a library wanting to purchase access to only a few electronic journals will be required, in practice, to purchase all of the publisher’s offerings. According to an e-mail from economist Mark McCabe to the author, “In the non-bundled environment, libraries have some limited ability to substitute away from very high cost/use titles. In the bundled environment, the willingness of libraries to drop entire bundles is obviously far less, reducing the opportunities for entry, etc.”
15
purchase a second product. A tying arrangement is said to be per se illegal where the plaintiff proves that (1) there are two distinct products, (2) the seller has required the buyer to purchase the tied product in order to obtain the tying product, (3) the seller has market power in the market for the tying product, and (4) the tying arrangement has a detrimental effect on competition in the market for the tied product, that is, there must be foreclosure of a substantial volume of commerce. From a prospective plaintiff’s point of view, an advantage of bringing a tying case is that tying is considered a per se violation, which does not require the same amount of proof that is required in most restraint of trade cases that are tried under the so-called Rule of Reason. This advantage has been melted away, to some extent, by various cases that have brought the proof requirements closer to those of a structured Rule of Reason case. Because tying can be either pro-competitive or anticompetitive, an efficiency defense is permitted. A classic tying case was U.S. v. Loew’s Inc .in 1962.16 The defendants held the rights to old films that the television networks wanted to use. Each of the defendant companies insisted on selling popular films only as part of a ‘package deal’ that included unpopular films that were not desired by the networks. For example, to obtain ‘The Man Who Came to Dinner,’ one network was required to purchase both ‘Gorilla Man’ and ‘Tugboat Annie Sails Again.’17 The Supreme Court held this to be illegal. In the 1982 Jefferson Parish case, the Supreme Court made tying more difficult to prove.
18
Dealing with the coercion requirement, the Court used fairly expansive
language, saying the tying product had to have “some special ability –usually called ‘market power’—to force a purchaser to do something that he would not do in a competitive market.” But in applying this, the Court seems to have limited its application to a showing of interbrand market power, examining whether the seller had a substantial
16
Supreme Court, 1962.
17
A.D. Neale, The Antitrust Laws of the U.S.A., 214-215 (Cambridge: Cambridge University Press, 1970).
18
Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2 (1984).
16
market share in the tying product.19 Jefferson Parish has been read to say that the tying product must have a minimum of 30% of the market in order to have coercive market power. In the Big Deal, there are arguably two distinct products, the printed subscription and the electronic data base. But here the tying analogy begins to become more difficult. Which is the tying product and which the tied? In most tying cases, the seller is coercing the buyer into purchasing a product he doesn’t really want and would not purchase in the absence of the tying arrangement. Here, it might be argued, the library actually wants both the printed subscriptions and the electronic data base. There is an option to buy products separately and in fact several large universities have exercised this option rather than submit to the Big Deal. In litigation, this fact would likely be used by defendant publishers to suggest strongly that the Big Deal is not coercive, even though the publisher offers terms that the Godfather might say are too good to be refused. There is also a question of whether the seller has market power in the tying product. In one sense, the publisher has a monopoly at each end of the deal, which is to say that no other publisher can offer precisely the same differentiated product because the specific content of the various titles is unique. In the Loew’s blockbooking case, the District Court had found that each copyrighted film blockbooked by the defendants for television use “was in itself a unique product.” By this line of reasoning, each title offered by the publisher may itself be a unique product with some degree of tying power. On the other hand, when the publisher makes a sales call at the library, it is in direct competition with other publishers who can provide alternative but similar content. It is difficult to say that Elsevier, with only about 25% of the market, measured by revenue, has monopoly power based on market share. Yet, it is possible that in more narrowly defined markets, the publisher really does have a market share that would satisfy the Jefferson Parish threshold.
19 See Lawrence A. Sullivan and Warren S. Grimes, The Law of Antitrust (St. Paul: West Group 2000) at 415.
17
Two further notes on this. First, market power is not always related to market share and where the power to control prices is shown directly, as it might be in this industry, it is not necessary either to fuss over market definition or to construct and rely on market shares. Unfortunately the precedents for this proposition have never been accepted by courts in mergers or tying arrangement cases. Second, anticompetitive effects can occur if sellers imposing the tie have a high collective market share.20 This seems to be the case in academic publishing. Whether there is case support to recognize this observation will require more legal research. I’ve been speaking in terms of the printed subscriptions being the tying product. Maybe the subject can be turned around. Suppose it can be established that the product that is most desired is the electronic data base. If a library has that, arguably it does not even need expensive hard-copy subscriptions. Does the publisher have market power over the electronic data base that contains all or a selection of its publications? The same questions about market definition seem to apply. In addition, exactly what is the market for electronic data bases? Who are the competitors in this market? What are the market shares? Assuming we can get satisfactory answers to these questions, an interesting issue arises. One of the problems with tying arrangements can be that the arrangement creates an informational void for the consumer, which can be exploited, particularly if the buyer purchases the tied product at a later date.21 In the Big Deal situation, the buyer is required to continue to subscribe to journals it may not even want and to pay prices that may escalate at the publisher’s will. During the period of the contract, market conditions may change and the buyer’s need for the tied product may change. This characteristic fits many of the cases in which the Supreme Court has held that a harmful tie may be present.22 20
Sullivan & Grimes at 405. The authors do not cite any cases to support this reasonable proposition.
21
See Sullivan & Grimes at 406-07: “Even if the buyer is concerned about the terms for the deferred purchase of the tied product, the buyer may not know how much (if any) of the tied product will be needed. Nor will the buyer be in a position to know what alternatives to the tied product might exist at the uncertain date in the future when the tied product is required.” 22
Sullivan & Grimes at 407 cite Kodak, 504 U.S. 451 (1992); Northern Pacific Railway Co. , 356 U.S. 1 (1958); International Salt Co. , 332 U.S. 392 (1947); and IBM Corp. , 298 U.S. 131 (1936).
18
There are three basic ways to tie products together. In some cases, it can be done through actual integration, as when Microsoft bundles the browser into the operating system. It can be done by contract. And it can be done through financial incentives, as in the Big Deal. The business world has become aware that exclusionary transactions do not have to be all-or-nothing in form. There exist a variety of transactions that are structured to avoid looking like they were intended to exclude competitors, but are actually intended to have the practical effect of exactly that. The “bundled rebate” is one of those strategies that has been generated with sufficient ambiguity that it may foreclose rivals while avoiding obvious antitrust liability. Whether the law has caught up with this reality of the marketplace happens to be one of the cutting edge questions in antitrust. Monopoly Maintenance and Bundled Rebates Monopoly maintenance is illegal under Section 2 of the Sherman Act. It requires two elements to be proved: (1) the possession of monopoly power and (2) the “willful acquisition or maintenance of that power as distinguished from growth or development as a consequence of a superior product.”23The traditional way to show monopoly power is through a high market share (usually 60 percent or more) in a well-defined market. Recall that the market leader, Elsevier, has only about 25 percent of the academic publishing market (not defined with precision). Edlin and Rubinfeld argue that there is now sufficient precedent to prove monopoly power directly by demonstrating that a company has been able to raise the price of its product substantially above competitive levels and to maintain that price increase for a substantial period of time.24 They are able to demonstrate that the price of commercial journals is much higher (200 to 400 percent higher) than for non-profit journals and they argue that this difference is not accounted for by quality differentials or by subsidization of the non-profits by their parent associations. Additionally, they conclude that Elsevier’s profit margins are high enough 23
U.S. v. Grinnell Corp., 384 U.S. 563, 570-71 (1966).
24
Edlin and Rubinfeld, op. cit.
19
to constitute evidence of monopoly power over prices. The law is still evolving as to whether this case would be enough to prove monopoly power in the absence of a monopolistic market share, but the logic is quite strong. The next question would be whether the bundled rebate strategy of excluding competitors from the market constitutes monopoly maintenance. Is Elsevier, for example, competing on the merits through high quality or low-cost service, or is its strategy preventing entry of superior journals? The facts to answer that central question would in all likelihood have to be developed within the context of litigation. Meanwhile there remains a certain vagueness about how the law should treat bundled rebates. A case that had the potential of clarifying the law on bundled rebates is 3M Company v. LePage’s Inc.25 Here is a statement of the basic facts, as taken from the U.S. government’s brief to the Supreme Court, filed May 28.
3M manufactures Scotch-brand tape and other products. Until the early 1990s, 3M had more than a 90% share of the United States market for transparent and invisible tape. Thereafter, 3M's share began to erode with the rise of office supply "superstores" (such as Staples and Office Depot) and the growth of mass merchandisers (such as Wal-Mart and Kmart), which sold products, including tape, under "private labels." LePage's expanded its tape line to include private label tape and, by 1992, LePage's had an 88% share of the growing private label market segment (but only 14.4% of the overall market). 3M reacted by entering the private label segment and by selling some tape under the "Highland" label… The alleged unlawful conduct included various "exclusive dealing arrangements" 3M secured through cash incentives, "bundled rebate" programs that "offered higher rebates when customers purchased products in a number of 3M's different product lines," and other conduct...
25
324 F.3d 141 (2003).
20
The challenged bundled rebate programs "offered discounts to certain customers conditioned on purchases spanning" multiple product lines, with the size of the rebate dependent on the customer's success in meeting 3M-established growth targets for the individual product lines. LePage's contended those bundled rebates helped 3M maintain its monopoly, because failure to meet the target for one product (such as tape) could cause a customer to lose rebates across multiple products. To make a purchase of LePage's private label tape financially attractive to a potential customer, LePage's alleged, it would not be sufficient to match 3M's price on similar tape. Rather, LePage's would have to reduce its private label tape price by an amount sufficient to compensate the purchaser for the loss of rebates based on the far larger volume of purchases the customer made on the full range of 3M products (including Scotch tape and non-tape products). 3M's strategy, LePage's alleged, was designed to forestall competition to its higher priced Scotch brand from private label tape. The Big Deal is a form of bundled rebate. When the library agrees to purchase both the printed subscriptions and the digital data base, the two products are bundled. The discount made available when the bundled package is purchased is functionally equivalent to a rebate. Here we must note several differences between the LePage complaint and the complaint that might be made by a small academic publisher who has been foreclosed by a Big Deal from an opportunity to sell to a library. First, 3M is a certified monopolist, based on large market share. As pointed out, the market share of any single academic publisher may not be in the same category, so market power will have to be demonstrated by focusing on the collective market share of publishers using the same tying device; or directly, for example by evidence of unusually high profits and/or unusually high prices. I also pointed out that these are not well-developed legal theories. Second, the reduction of prices implicit in 3M’s bundled rebates has a clear benefit to purchasers, which might be offered up in defense. It is not clear that a publisher’s rebate provides any direct benefit to the library or creates any significant efficiencies. Certainly, it can be argued by a publisher that the Big Deal makes a discount
21
available, but that is only a shift of money from one pocket to another and it is not clear that there is any recognizable efficiency in this that would not have occurred had the publisher sold the products separately at discounted prices. Thus, one could debate whether the analogy is close, but at least the LePage’s case allows the law to address the circumstances under which a bundled rebate strategy is illegal. There is a reason why I have not yet disclosed whether LePage’s prevailed. In District Court, the opponent of the bundling rebate, LePage’s, won a large jury verdict. A panel of the Court of Appeals reversed. The full appellate court reheard the case en banc and a divided Court held for LePage’s. 3M then appealed to the Supreme Court on the question of whether Section 2 (monopolization) of the Sherman Act was violated. As often happens, the Supreme Court asked the Solicitor General for advice on whether to grant certiorari. The government’s long-awaited answer, reflecting FTC and DOJ views, urged against granting certiorari. The Supreme Court took this advice.26 The LePage victory stands, but a good bit of vagueness surrounds the question of what it stands for. The government’s brief arguing that the Supreme Court should not decide the LePage case reflects concern that the LePage appellate opinion was wrongly determined and poorly explained, in that it may encourage plaintiffs to bring similar suits (something the federal government enforcers believe would be bad), but it also takes the strategic position that it would be premature for the highest court to take a position on bundled rebates. Here is what the government said: The court of appeals' decision in this case addresses the application of Section 2 of the Sherman Act to the business practice of "bundled rebates." The en banc court of appeals rejected petitioner's primary contention that this Court's decision in Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993), precludes Section 2 liability for the bundled rebates at issue here because they did not result in below-cost pricing. That ruling does not conflict with the decisions of any other court of appeals. While it would be desirable to provide the business community, consumers, and the lower courts with additional guidance on the application of Section 2 to bundled rebates, this case does not provide a
22
suitable vehicle for providing such guidance. The court of appeals was unclear as to what aspect of bundled rebates constituted exclusionary conduct, and neither it nor other courts have definitively resolved what legal principles and economic analyses should control. In addition, there is substantial uncertainty in the record below concerning facts that may be significant. Because the issues here are novel and difficult, and because petitioner fails to demonstrate an urgent need justifying this Court's immediate intervention, the Court should deny the petition for a writ of certiorari and allow the lower courts an opportunity to refine and clarify the application of Section 2 to this particular business practice.
Looking at Horizontal and Vertical Competition Together We’ve focused separately on the horizontal issues presented by mergers of academic publishers and seen that reliance has been placed on the standard market definition approach in which the only anticompetitive potential is found in overlapping titles. This myopic but common approach may result in some divestitures but it does not stop the merger wave. I’ve suggested that two alternative approaches are worth pursuing when the next merger occurs: (a) expand the market definition by viewing the market from the buyer’s perspective, which yields a more realistic picture of how this market works; and (b) develop the case for protecting consumer choice by stressing the increased ability of publishers to exercise buyer power in the acquisition of content. I do not suggest that either of these approaches has a high probability of succeeding in the current federal environment, but on the whole I believe they are sufficiently promising to merit the devotion of further resources. It is also possible that one or more States would take up the cudgels, in view of the fact that many States operate libraries that will likely be affected by further concentration. The States are often more receptive than federal law enforcement agencies to the types of approaches we are discussing, and especially when it is their purchasing budget that is in question.
26
The Supreme Court denied certiorari in 3M Company v. LePage’s, Inc., on June 30, 2004.
23
We’ve also focused separately on the vertical issues presented by Big Deal bundling and seen that the law of bundling rebates is unsettled but offers a promising avenue for attacking the current marketplace constriction on anti-tying grounds. Again, creative thinking on the part of the libraries’ lawyers and economists will be needed, but the situation is certainly not hopeless. Are the horizontal and vertical issues related? It seems to me that they are. Without the consolidation of publishing that is being accomplished through the horizontal merger process, no publisher would have sufficient market power to impose the Big Deal on its buyers. It is only when there are a small number of very powerful sellers that this sort of arrangement can be pushed. If librarians had more choices, they would have more leverage with which to bargain with any individual publisher. If this connection can be developed and supported by evidence, it can be used in argument against the next big merger, either to stop the merger from being consummated or to insist that a condition be that the merged company not engage in bundling rebates, so as not to further increase entry barriers or reduce the options available to researchers and writers in the distribution of their products. In concluding, let me point out the close relationship between what occurs in the publishing marketplace and the civil liberties of Americans. It is sometimes said that the Sherman Act is the Magna Carta of our economy, a part of the framework that has a majestic, almost constitutional presence. Indeed, antitrust fits nicely with constitutional concepts like division of powers, checks and balances, and limited government. When we come to markets in which information is produced and distributed, we are even closer to the core of our Constitution, because democracy depends on the free flow of information, which is specifically protected by the First Amendment. It is potentially dangerous both to the First Amendment and to the marketplace when academic publishing becomes highly concentrated. The fact that it is journals instead of shoes that we are talking about
24
is enormously important, but I have to tell you that in the antitrust law, this difference is ignored.27 Between the merger wave and the invention of the Big Deal, not only the nation but the English-speaking world seems to be headed for that dangerous territory in which a small number of individuals, working through international corporations, may gain the power to control important aspects of the production and distribution of critically important information. We have an obligation to stop this movement.
27
For an interesting treatment of the relationship between competition and free speech, see Maurice E. Stucke and Allen P. Grunes, “Antitrust and the Marketplace of Ideas,” 69 Antitrust L.J.249 (2001).
25